Spiral wound modules and spacers

Abstract

The operation of spiral wound modules in industrial plants is affected by many parameters, including the operating conditions, the arrangements of the spiral wound modules in arrays and the design of the spiral wound module itself. This paper reviews techniques and approaches for the analysis and optimisation of the performance of spiral wound modules. The analysis of the design of spiral wound modules and arrays with a combination of experimental and numerical techniques can help to identify the optimal array arrangements, module geometry and spacer design for specific applications. The onset of fouling as characterised by the decisive point of the onset of fouling known as the critical flux is discussed in the context of minimisation of fouling within modules and arrays.

Introduction

Membrane processes are now widely considered as economical alternatives to conventional separation processes. Reverse osmosis (RO), nanofiltration (NF), ultrafiltration (UF) and microfiltration (MF) have become standard unit operations. Commercially available modules include spiral wound, hollow fibre, tubular and plate-and-frame modules. Amongst these, the hollow fibre modules and spiral wound modules (SWMs) are the most common, due to their high membrane area to volume ratio. Even though the hollow fibre modules have a higher packing density than the SWM, the latter are often preferred in industry because they offer a good balance between ease of operation, fouling control, permeation rate and packing density. SWMs are widely used for commercial applications ranging from RO to UF. The applications of SWMs include desalination, water treatment, water reclamation, treatment of industrial waste water, product treatment in the dairy industry and recovery of valuable products in the pharmaceutical industry [1], [2], [3]. The major problems for an SWM are concentration polarization, fouling and high pressure loss.

The performance of an SWM is affected by many factors:

• SWM leaf geometry, such as number of leaves, leaf length and width, feed and permeate channel heights,

• Spacers, which greatly affect local mixing, mass transfer and pressure loss,

• Fouling propensity and cleaning ability,

• Operating conditions, such as feed pre-treatment, feed concentration, feed pressure and permeate recovery.

This paper reviews and analyses the effects of the above parameters on SWM design and performance.

A schematic diagram of an SWM is shown in Fig. 1. The major components of an SWM are the membrane, the feed and permeate channels, spacers which keep the membrane leaves apart, the permeate tube and the membrane housing [3], [4]. The feed channel spacer also enhances mass transfer near the membrane but inevitably increases pressure loss along the membrane leaf [5]. Membrane sheets with the spacers in between are glued together on three sides to form a leaf and multiple leaves are rolled up around the permeate tube to create the feed and permeate channels of the SWM. A pressurised module housing holds the membrane leaves in place to prevent unwinding. Usually three or more modules are connected in series in the housing.

The geometry of an SWM is described by the number of leaves, N_L, the leaf length, L, and leaf width, W, of each membrane leaf, the feed channel height, h_f, and permeate channel height, h_p. The channel heights are defined by the feed and permeate spacer heights. The spacers themselves are characterised by the mesh length (distance between filaments), filament diameter, orientation of the filaments, hydraulic diameter and voidage [5], [6].

The feed solution flows in an axial direction parallel to the permeate tube through the feed channel. The solvent passes through the membrane and flows as permeate spirally along the curved permeate channel until it is collected in the permeate tube. Fig. 2 illustrates in more detail a flat membrane leaf of an SWM as it would appear before winding around the central tube. The feed solution flowing along the x-axis enters at x = 0 and exits at x = L. The permeate flows along the y-axis. Some permeate flow starts at y = 0 which is the closed end of the permeate channel and all the permeate exits the membrane leaf at the permeate collector tube at y = W. The operating conditions which are adjustable are the feed inlet flow rate, feed concentration, feed pressure and the permeate tube pressure.

The easiest experimental technique for studying fluid flow across a leaf of an SWM utilizes a flat channel test cell. Such a test cell provides a simple but effective method of studying concentration polarization, flux and pressure loss under various operating conditions, such as feed pressure, feed velocity or feed concentration [5], [7], [8], [9], [10], [11]. In addition, the effect of feed spacers on flux enhancement and pressure loss can be tested [5], [12], [13], [14], [15], [16], [17], [18]. However, due to a compression of the spacer mesh which occurs in SWM during the wrapping, test cell results may be limited. A theoretical model can provide further insight into the local flow patterns in spacer-filled channels while modelling of flow across the leaves together with data from module autopsies can help develop understanding of the relationship between flux patterns and fouling within the SWM.